URGENT help for RESTORE

rmarzullo · #1 (**permalink**) 07-12-07, 13:06

I'm running a DB2 7.1 (old) and I have 4 servers, 2 are production and 2 QA.One of QA servers, which I had 2 databses, I used to restore online backups evryday without any problems...
One day this restore failed, due to another DBA issued a "force all" while my restore was running. I never was able to restore successfuly or rollforward again. I try dropping everything and restore...didn't work because one of the DB has a different code set from the backup, so I created a empty db with same name and ran the same restore script..but now I'm getting those errors on DB2DIAG.LOG:

2007-07-12-12.56.40.550559 Instance:db2owner Node:000
PID:18112(db2bm.39842.0) Appid:none
buffer_pool_services sqlbSMSDoContainerOp Probe:835
DIA9999E An internal error occurred. Report the following error code :
"FFFFC11E".

2007-07-12-12.56.40.575842 Instance:db2owner Node:000
PID:18112(db2bm.39842.0) Appid:none
buffer_pool_services sqlbSMSDoContainerOp Probe:835

Error checking container 0 (/db2owner/db2owner/NODE0000/SQL00001/SQLT0000.0) for tbsp 0. Rc = FFFFC11E

2007-07-12-12.56.40.652896 Instance:db2owner Node:000
PID:18112(db2bm.39842.0) Appid:none
oper_system_services sqloopenp Probe:36

errno: 0000 001a ....

2007-07-12-12.56.42.710182 Instance:db2owner Node:000
PID:18112(db2bm.39842.0) Appid:none
oper_system_services sqloopenp Probe:36

errno: 0000 001a ....

2007-07-12-12.56.42.825035 Instance:db2owner Node:000
PID:18112(db2bm.39842.0) Appid:none
oper_system_services sqloopenp Probe:36

errno: 0000 001a ....

2007-07-12-12.56.43.067187 Instance:db2owner Node:000
PID:18112(db2bm.39842.0) Appid:none
buffer_pool_services sqlbDMSAcquireContainer Probe:105

Container '/dev/remselv01' is different from specified size.
User: 983040 Kb OS: 1966080 Kb. Extra storage is wasted.

2007-07-12-12.56.43.657582 Instance:db2owner Node:000
PID:18112(db2bm.39842.0) Appid:none
buffer_pool_services sqlbDMSAcquireContainer Probe:105

Container '/dev/remselv02' is different from specified size.
User: 983040 Kb OS: 1966080 Kb. Extra storage is wasted.

-----------------
I tried to db2untag those containers and run it again...but still got the same errors...

I need help URGENT please!

Thanks

stolze · #2 (**permalink**) 07-13-07, 04:35

Interpreting the first RC says:

Code:

$ db2diag -rc FFFFC11E

Input ECF string 'FFFFC11E' parsed as 0xFFFFC11E (-16098).
NOTE: ../sqz/sqlzwhatisrc.C:
        V7 input ZRC 0xFFFFC11E (-16098) may translate to V8 ZRC value of 0x8402001E (-2080243682)

ZRC value to map: 0x8402001E (-2080243682)
        V7 Equivalent ZRC value: 0xFFFFC11E (-16098)

ZRC class :
        Non-Critical Media Error (Class Index: 4)
Component:
        SQLB ; buffer pool services (Component Index: 2)
Reason Code:
        30 (0x001E)

Identifer:
        SQLB_CONTAINER_NOT_ACCESSIBLE
Identifer (without component):
        SQLZ_RC_CNACCESS

Description:
        Container not accessible

Associated information:
        Sqlcode -293
SQL0293N  Error accessing a table space container.

        Number of sqlca tokens : 0
        Diaglog message number: 1

In other words, the container /db2owner/db2owner/NODE0000/SQL00001/SQLT0000.0 cannot be accessed. Further, looking at errno 0x1A, which is encountered by sqloopenp (file open), in /usr/include/asm-generic/errno-base.h on my Linux system tells me that this error is ETXTBSY and means "Text file busy". You may want to double-check the meaning of errno 0x1A on your platform (you haven't told us that).

Then you should check if the container exists and if there are any interesting things to be found (maybe it is read-only, maybe a link pointing to Nirvana, maybe the directories leading to the container don't exist, maybe something else.

Also, you seem to have some raw containsers (/dev/remserv0*). DB2 notices that the raw disks have a larger size than the container size. You should investigate why this was configured in such an unusual way and fix it, i.e. don't waste the disk space. But those messages are not really an immediate problem, I'd say.

guyprzytula · #3 (**permalink**) 07-13-07, 04:37

I would execute drop db and remove path /db2owner/db2owner/NODE0000 and below. then try restore...

rmarzullo · #4 (**permalink**) 07-13-07, 08:55

I'd like to thank you both of you Knut Stolze and Guy Przytula.
I was able to restore the databases on thise server (AIX 4.3).

The whole mess started when a restore failed (someone kill the processe accidentally) and even after trying couple restores, I couldn't rollforward due to system error. So I decided to drop everything and start from scratch, but .... the sqldbdir file was damaged. So I deleted all files under NODE0000, untag (db2untag) all raw containers that were associated to both databases, recreate empty 2 databases and restored then. My problem is sometimes I got confused with some messages that are not meaningful.

Again thanks for your help.

Rmarzullo

Internet Services

Web Development

Digital Marketing

Consumer Tech